Intra-specific variation and allometry of the skull of Late Cretaceous side-necked turtle Bauruemys elegans (Pleurodira, Podocnemididae) and how to deal with morphometric data in fossil vertebrates Thiago F. Mariani1,2 and Pedro S.R. Romano1 1 Deparatamento de Biologia Animal, Universidade Federal de Vicosa,¸ Vicosa,¸ MG, Brazil 2 Current affiliation: Departamento de Geologia e Paleontologia, Museu Nacional, Universidade Federal do Rio de Janeiro, Rio de Janeiro, RJ, Brazil ABSTRACT Background. Previous quantitative studies on Bauruemys elegans (Suárez, 1969) shell variation, as well as the taphonomic interpretation of its type locality, have suggested that all specimens collected in this locality may have belonged to the same population. We rely on this hypothesis in a morphometric study of the skull. Also, we tentatively assessed the eating preference habits differentiation that might be explained as due to ontogenetic changes. Methods. We carried out an ANOVA testing 29 linear measurements from 21 skulls of B. elegans taken by using a caliper and through images, using the ImageJ software. First, a Principal Components Analysis (PCA) was performed with 27 measurements (excluding total length and width characters; =raw data) in order to visualize the scatter plots based on the form variance only. Then, a second PCA was carried out using ratios of length and width of each original measurement to assess shape variation among individuals. Finally, original measurements were log-transformed to describe allometries over ontogeny. Results. No statistical differences were found between caliper and ImageJ measure- Submitted 5 June 2016 Accepted 9 December 2016 ments. The first three PCs of the PCA with raw data comprised 70.2% of the variance. Published 11 April 2017 PC1 was related to size variation and all others related to shape variation. Two Corresponding author specimens plotted outside the 95% ellipse in PC1∼PC2 axes. The first three PCs of Thiago F. Mariani, the PCA with ratios comprised 64% of the variance. When considering PC1∼PC2, all [email protected] specimens plotted inside the 95% ellipse. In allometric analysis, five measurements were Academic editor positively allometric, 19 were negatively allometric and three represented enantiometric Laura Wilson allometry. Many bones of the posterior and the lateral emarginations lengthen due to Additional Information and increasing size, while jugal and the quadratojugal decrease in width. Declarations can be found on Discussion. ImageJ is useful in replacing caliper since there was no statistical dif- page 33 ferences. Yet iterative imputation is more appropriate to deal with missing data in DOI 10.7717/peerj.2890 PCA. Some specimens show small differences in form and shape. Form differences Copyright were interpreted as occuring due to ontogeny, whereas shape differences are related 2017 Mariani and Romano to feeding changes during growth. Moreover, all outlier specimens are crushed and/or distorted, thus the form/shape differences may be partially due to taphonomy. The Distributed under Creative Commons CC-BY 4.0 allometric lengthening of the parietal, quadrate, squamosal, maxilla, associated with the OPEN ACCESS How to cite this article Mariani and Romano (2017), Intra-specific variation and allometry of the skull of Late Cretaceous side-necked turtle Bauruemys elegans (Pleurodira, Podocnemididae) and how to deal with morphometric data in fossil vertebrates. PeerJ 5:e2890; DOI 10.7717/peerj.2890 narrowing of jugal and quadratojugal may be related to changes in feeding habit between different stages of development. This change in shape might represent a progressive skull stretching and enlargement of posterior and lateral emargination during ontogeny, and consequently, the increment of the feeding-apparatus musculature. Smaller individuals may have fed on softer diet, whereas larger ones probably have had a harder diet, as seen in some living species of Podocnemis. We conclude that the skull variation might be related to differences in feeding habits over ontogeny in B. elegans. Subjects Paleontology, Zoology Keywords Morphometry, Ontogeny, Pelomedusoides, Bauru group, Testudines, Principal components analysis, Missing data, Feeding habits INTRODUCTION Principal Component Analysis and fossil sampling bias Paleontological data are intrinsically scarce (Strauss, Atanassov & Oliveira, 2003; Hammer & Harper, 2006), leading to incomplete data sampling. This limitation impacts several approaches in paleontological studies, especially inter-specific variation analyses. Although there are some methods proposed to deal with missing entries in fossil quantitative datasets (e.g., Norell & Wheeler, 2003; Strauss, Atanassov & Oliveira, 2003), sometimes the study relies on an exploratory evaluation of general structure and Principal Component Analysis (PCA) is commonly used for this purpose. PCA is a method to ordinate multivariate data. Its aim is to identify the variables that account for the majority of the variance within a multivariate matrix, by means of linear combinations of all variables, which are converted into components that are independent of each other (Strauss, Atanassov & Oliveira, 2003; Hammer, Harper & Ryan, 2001). Hence, PCA summarizes a large amount of the variance contained in the data (Krzanowski, 1979; Hammer, Harper & Ryan, 2001). It thus reduces a multidimensional space into fewer components which retain the majority of the variance of a given sample (Jolicoeur & Mosimann, 1960; Peres-Neto, Jackson & Somers, 2003), and is therefore an useful tool for exploring large, complex data sets, being largely applied to both extant and turtles (e.g., Jolicoeur & Mosimann, 1960; Claude et al., 2004; Depecker et al., 2005; Depecker et al., 2006; Werneburg et al., 2014; Ferreira et al., 2015. Case-study Skull variation The skull is one of the most variable structures in vertebrates because it concentrates several sensory organs, the brain, and the beginning of the respiratory and digestory systems, including chewing muscles (Smith, 1993). Consequently, the skull is the body part with more phenotypes used in vertebrate cladistic analysis (Rieppel, 1993), as seen in turtles, in which most cladistic analysis rely mainly on cranial characters (e.g., Gaffney, Meylan & Wyss, 1991; De La Fuente, 2003; Gaffney, Tong & Meylan, 2006; Gaffney et al., 2011; Joyce, 2007; Joyce & Lyson, 2010; Sterli et al., 2010; Sterli & De La Fuente, 2011; Anquetin, 2012; Rabi et al., 2013; Romano et al., 2014; Ferreira et al., 2015; Romano, 2016). Despite Mariani and Romano (2017), PeerJ, DOI 10.7717/peerj.2890 2/40 that, most skull materials found in paleontological record of turtles are rare and/or damaged due to the fossilization process bias, not allowing intraspecific comparisons or ontogenetic inferences on most fossil turtle species known. Some exceptions are found in Sánchez-Villagra & Winkler (2006) and Ferreira et al. (2015), who performed interspecific comparisons among fossil turtle taxa using skull material. Bauruemys taxonomy Bauruemys elegans (Suárez, 1969) is a Late Cretaceous freshwater side-necked turtle found at the Pirapozinho site (Suárez, 2002), in western São Paulo state. This species was originally described as Podocnemis in three different communications by Suárez (1969a), Suárez (1969b) and Suárez (1969c) and such recognition was based on the overall similarities of the skull and shell to this living genus, a common practice that time. Other South American Cretaceous side-necked turtles were initially identified as Podocnemis as well, such as the nomina dubia ``Roxochelys'' harrisi (Pacheco, 1913) and ``Bauruemys'' brasiliensis (Staeche, 1937) and the incertae sedis ``Podocnemis'' argentinensis (Cattoi & Freiberg, 1958) (see Romano et al., 2013 for a revision on Bauru Group species). Kischlat (1994) was the first to point out that all Podocnemis reported from the Cretaceous were doubtful and proposed a new genus, Bauruemys, to include B. elegans and, tentatively, B. brasiliensis. His conclusion was based on similarities of the plastron of both species, besides the analysis of cranial features in B. elegans. More recently, Romano et al. (2013) confirmed the recognition of B. brasiliensis as Bauruemys, but considering this species as nomem dubium. Kischlat (1994) and Kischlat, Barbarena & Timm (1994) also pointed out that B. elegans could belong to Podocnemididae, but they did not test their hypothesis. Romano & Azevedo (2006) were the first to carry out a cladistic analysis to access the phylogenetic position of Bauruemys, placing it as a stem-Podocnemididae, i.e., the sister group of all other Podocnemididae, which was consistently confirmed by subsequent analyses with more podocnemidid species included as terminals (Fran¸ca & Langer, 2005; Gaffney et al., 2011; Oliveira, 2011; Cadena, Bloch & Jaramillo, 2012). Geological settings and taphonomic context of the Tartaruguito site The Pirapozinho site, long ago known as ``Tartaruguito'' and formally assigned as such by Romano & Azevedo (2007) and Gaffney et al. (2011), is an Upper Cretaceous outcrop from the Presidente Prudente Formation, Bauru Basin (Geology sensu Fernandes & Coimbra, 2000). It is located in Pirapozinho municipality, São Paulo State, Brazil (Fig. 1). The ``Tartaruguito'' name, which means ``turtle in rock'' (tartaruga, from Portuguese, turtle; ito, from Latin, rock. More precisely, ``ito'' is the Portuguese suffix correspondent to ``ite,'' the suffix used in geology to forms the names of rocks and minerals.),